Title of the Project Machine learning Performance Improvement Using AI based Data Centric Approach

Project Title

Title of the Project Machine learning Performance Improvement Using AI based Data Centric Approach

Project Area of Specialization Software EngineeringProject Summary

AI based applications is facing different problems regarding accuracy of data. A lot of time is wasted on improving model while the data is not handled in a better way. As machine learning becomes the new software, collecting data and improving its quality will only become more critical for deep learning. We covered four major topics (data collection, data cleaning and validation, robust model training, and fair model training), which have been studied by different communities, but need to be used together. 

Project Objectives

To increase efficiency and effectiveness of data by data-centric approach. 

 To improve overall quality of data by using different approaches of data quality improvement. 

 To improve performance and accuracy of AI systems by enhancing the datasets.

Project Implementation Method

Improves the prediction quality by using New method( Data centric approach)

Change the accuracy of dataset by 

Data Preprocessing of diabetes data

Data quality assessment

Data cleaning

Handling missing values

Data transformation

Data Augmentation of diabetes data

Increase the amount of data 

Imcrease Accuracy and performance of dataset

Data Preprocessing of MNIST Dataset

Data Programming of MNIST DIGIT Dataset 

          Using GAN

Benefits of the Project

It will increase the amount of data artificially by using data augmentation and also will reduce overfitting issue.

It will preprocess the data .

It will help on deep learning of data  by cleaning daat and also increasing amount of data systematically.

Technical Details of Final Deliverable

The final deliverable will be in the form of webapp that shows tha improved accuracy of the datasets will approved that working on datasets is far better than improving tha models because the AI basically consist of 80Úta and the remaining code . So it's. Better to work on data rather than code  or models.

Final Deliverable of the Project Software SystemCore Industry ITOther IndustriesCore Technology Artificial Intelligence(AI)Other TechnologiesSustainable Development GoalsRequired Resources

Item Name	Type	No. of Units	Per Unit Cost (in Rs)	Total (in Rs)
			Total in (Rs)	33000
Online AI tools	Equipment	1	18000	18000
Website hosting	Equipment	1	10000	10000
Misclaneous header	Miscellaneous	1	5000	5000